Research On Electronic Business Intelligence Recommend System Based On Rough Set

Rough set theory (Rough Set) is the tool of mathematics proposed by professor Pawlak whois a mathematician from Poland in1982which can quantitatively analyze and deal withimprecise, inconsistent, incomplete information and knowledge, commonly used to deal withfuzzy and uncertain problems. Data mining is knowledge discovery in database (KnowledgeDiscover Database, KDD), and commercial data mining in business intelligence research is aprocess that extracts implicit information with potential value for the invoicing of commercialretail enterprises from database. Rough set data mining system (RSDMS) can be used for datapreprocessing, eliminating redundant attributes, extracting decision rules, reducing attribute setin the condition of not affecting the decision-making efficiency. Through applying SQL to thereduction process of seeking information entropy，correlation reduction algorithm can improveits efficiency by shortening the execution time. Finding frequent itemsets algorithm is one of themain content of association rules, this paper proposed a kind of improved Apriori algorithmbased on literature [44], we can easily get the support of itemsets count based on the matrix, theimproved algorithm does not require as frequently as Apriori algorithm when scanningtransaction database which can reduce the scanning times for transaction database, and at thesame time, it doesn’t need to obtain each set of corresponding sub matrix as the algorithm inreference [44] which can reduce the complexity of algorithm. On the basis of previous researchresults, this paper do research for the attribute reduction algorithm of rough set and itsapplication, commercial data mining, intelligent recommendation for electronic commerce andother aspects, the main contents are summarized as follows:(1) Did general research on basic theory knowledge of rough sets, introduced thedevelopment course and the present situation at home and abroad in detail, analyzed the attributereduction algorithm of rough set, and improved the algorithm using attribute reduction algorithmbased on SQL in condition of streamlining set of condition attributes.(2) Introduced data mining, basic theory and application of commercial data mining andproblems in application. Did general research and analysis on discretization and completion ofdata, proposed a kind of improved Apriori algorithm based on relevant literature.(3) Introduced the development course and the present situation at home and abroad ofelectronic commerce, analyzed the process and problems we should pay attention to during itsrunning process of electronic commerce, and did research on electronic commerce intelligent recommendation system.(4) Analyzed and dealt relational data of electronic commerce Website, did something aboutdata mining making use of improved attribute reduction algorithm of rough set and improvedalgorithm of Apriori, using the two improved algorithms in combination as a scheme inrecommendation system，and then comparised the efficiency of recommendation system bydoing experimental analysis for related data.